The ion beam sputtering coating process constitutes a complex system with multivariate, nonlinear, and strongly coupled dynamics, presenting significant challenges to the precision and intelligence of equipment control methods. This study proposes a dynamic closed-loop control approach based on a digital twin framework. A high-fidelity digital twin model of the ion beam coating equipment was developed, integrating both physical characteristics and real-time operational data. Centered on this twin model, a ”virtual-first, physical-execution, feedback-optimization” control architecture was designed: control commands are first simulated, validated, and optimized within the digital twin for outcome prediction before being deployed to the physical equipment for execution. Meanwhile, real- time production data are compared with the twin's predictions, and intelligent algorithms dynamically adjust both the control parameters and the twin model itself, establishing a closed-loop workflow that encompasses perception, decision-making, execution, and optimization. Furthermore, this study introduces real-time in-situ monitoring technology—an optical monitoring system—that dynamically tracks spectral transmittance during film growth. Based on the actual deposition data, online redesign and adaptive adjustments are made to the film stack design. Finally, experimental results from the automated coating of a specific high-pass filter demonstrate the effectiveness of the proposed method.
Pei et al. (2026) studied this question.